Human Metabolism Emits 2.05 Billion Tonnes of CO₂ Annually, Study Finds
A University of Bath‑led modelling study published in The Lancet Planetary Health estimates that human metabolic carbon dioxide rose 74 % between 1985 and 2015, reaching 2.05 billion metric tons per year. The research, led by Professor Javier Gonzalez of the Centre for Exercise, Nutrition and Metabolism, calculated the rise from 1.18 billion tons in 1985 to 2.05 billion tons in 2015, and attributes more than 60 % of the increase to global population growth.

The authors stress that metabolic CO₂ differs fundamentally from fossil‑fuel emissions. Carbon exhaled by people originates from recent plant photosynthesis that captured atmospheric CO₂ in food, whereas burning coal, oil or gas releases carbon stored underground for millions of years. This distinction means that while human respiration recycles carbon, it does not add net carbon to the atmosphere.
To derive the global figure, the team compiled thousands of real‑world measurements of CO₂ output across the lifespan—from infancy to old age—and paired these data with worldwide population and average body‑weight statistics. The analysis showed that per‑person metabolic emissions have remained relatively stable; the dominant driver of the total increase is the rise in the number of people, especially in densely populated urban areas.
The study’s authors say the new estimate can sharpen carbon‑budget models and improve the accuracy of climate‑policy scenarios that rely on precise source accounting. Gonzalez cautions that the findings are not a call to curb breathing or alter diets, but rather a tool for scientists assessing the effectiveness of carbon‑reduction programmes. The full paper, “Accounting for global carbon dioxide production from human metabolism between 1985 and 2015,” is available via DOI 10.1016/j.lanplh.2026.101546.
Editor's note: The article captures all essential data points, including the study's specific findings, methodology, and context, while maintaining professional tone.
AI-generated and fact-checked against the original report; claims the gate cannot verify are held back.